Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Strategies & Market Trends : Gorilla and King Portfolio Candidates -- Ignore unavailable to you. Want to Upgrade?

To: Mannie who wrote (25128)	5/21/2000 11:47:00 PM
From: buck	Read Replies (1) \| Respond to of 54805

<expound on Infiniband> Briefly, Infiniband continues the at-least-a-decade-long trend away from shared bus architectures to switched architectures. This change is now occurring inside of the server and workstation itself. The PCI bus, exactly what you think it is if you've bought a PC in the last few years, is being replaced by a cell- or frame-switched fabric. Bus-based architectures require that the bandwidth available on a bus be shared amongst all those entities that might access it, usually in order to move information around in a server. Say, for example, moving an information packet from memory to an Ethernet card for transmission on a network, or from a microprocessor to memory to store the result of a calculation. It is not unusual for an entity to have to wait for the bus to be free before transmitting information to another entity. Switched fabrics allow all entities to have full access at full speed to all other entities all the time. That's simplistic, but as deep as I feel I should go right now. For example, it allows memory to open a pipe for transmitting information to the ethernet card and to the microprocessor at the same time. An added bonus is that neither transmission interferes with the other, as it might in a shared-bus system. The potential is that it will boost transfer speeds and volumes within a server and potentially amongst servers by a factor of ten, thus removing the latest bottleneck in high-speed computing. At which point, we will find out what the next bottleneck is, and possibly, the next Gorilla. I hope this helps. If not, let me know. buck